This application describes research and development in biomedical computing in the areas of biomedical imaging, neurobiology and cognitive science, as described in NIH solicitation PAR-03-119. The objective is to develop novel computational technologies and methods for real time functional magnetic resonance imaging (rtfMRI), and using rtfMRI to direct behavior. This research will test the hypothesis that using this technology, it is possible to train subjects to activate spatially localized brain regions using real time fMRI, and that this will impact cognitive and behavioral processes, including the experience of pain. Initial pilot experiments suggest that in the near-term this technology will allow conceptually novel approaches to brain research, and will improve the speed and quality of fMRI data. Longer term, this technology may find important clinical applications in the diagnosis and rehabilitative treatment of central nervous system disease, with an initial focus on chronic pain. This grant will extend development beyond successful completion of stated phase I objectives, and extend research beyond preliminary data showing significant reductions in acute and chronic pain using this method. Specific Aims include: 1) Create production version of real time fMRI control software. 2) In normal volunteer subjects, measure enhancement of brain activation using real time fMRI-based training and thermally induced pain perception. 3) In chronic pain patients, measure enhancement of brain activation using rtfMRI-based training and pain chronic ratings.